NSTX TF Flag Joint Design Review SUMMARY C Neumeyer 8/7/3
NSTX TF Flag JointDesign Review
SUMMARY
C Neumeyer8/7/3
Topics
• Performance• Design Margins• Design Improvement Summary• Conclusions
PERFORMANCEFEA Element Pressures and Resistivities
Red cells > 1-in2
Temperatures Well Below Limit of 120oC
Contact Region
Max Temperature of 94oC Occurs Just After EOFT
• Tflat = 0.7sec (vs. 0.6 req’t)• Bolt Holes not exactly modeled (+10oC)• OH constant at max current (-TBDoC)• Insignificant change from constant resistivity simulation
OFF-NORMAL CASE: 3000lbf PRELOAD
Normal
• Peak Temperature ≈ 3oC Higher
• Temperature Distribution Different
• Current Redistribution Beneficial
Notes:1) Held SOFT pressure conditions after
SOFT due to lack of EOFT data2) Color scales different
Off-Normal
DESIGN MARGIN METHODOLOGY• State Safety Factor against Yield (as applicable) and against Failure
• Flag bolts, Collar bolts, Box bolts– Compare load to published material data for yield and ultimate tensile
• Threaded inserts and Shear Shoe bolts– Compare load to yield and failure data* from pullout tests– Failure data based on 3 statistics– Yield based on stress/strain test curves , average yield/failure ratio
• Wet lay-up– Compare load to failure data* from shear tests at 1ksi compression– Failure data based on 3 statistics
• FEA-derived loads– Use worst load from all “nominal” cases: =0.2,0.4, T= SOFT,EOFT,EOP
*Note: Data from uncycled samples
SHEAR TEST DATA INTERPRETATION
€
ˆ σ failure=kPA σ 0 −3σ +kNσN[ ]FEA peak/avg = 1.65
Shear Strength vs. Normal Compression
0
1000
2000
3000
4000
5000
6000
7000
8000
0 500 1000 1500 2000
Normal Compression (psi)
Shear Stress to Failure (psi)
Test DataFitFit-3*sigmaKPA*(Fit-3*sigma)
TEST DATA
Type Samples Avg 3 Avg-3Yield/Failure
Inert Pull-out in Copper 7 12714 2898 9817 0.84Bolt Pull-out in Copper 8 12658 2641 10016 0.82
et Lay-Up Shear 9 2486 845
0=2708 1000=4990
DESIGN MARGINSComponent Material
Max Load (lbf)
Max Stress (psi)
Yield Stress (psi)
Safety Factor on
Yield
Failure Stress (psi)
Safety Factor on
FailureFlag Studs Inconel 718 5755 74268 185000 2.49 210000 2.83
Insert pullout Copper 5755 13573 19412 1.43 23152 1.71Shear Shoe
Bolt A286 5279 68125 102000 1.50 146000 2.14Shear Shoe Bolt pullout Copper 5279 11971 18701 1.56 22713 1.90
Wet Lay-up Peak Shear
(Bond Plane) Hysol 123848 1500 4990 3.26
Wet Lay-up Peak Shear (Combined) Hysol n.a. 3000 4990 1.66Collar Bolts A286 9084 64018 102000 1.59 146000 2.28Box Bolts 304SS 5068 19800 31200 1.58 73200 4.36Box Bolt Pullout 304SS 5068 7303 15600 2.14 36600 5.01
Box Friction 304/304 4700 2.52
DESIGN IMPROVEMENTSFeature Old Design New DesignHub Stiffness Not adequate; lacking stiff linkages between disks
because flags could slide w.r.t. disksVery stiff. Boxes form webs with disks like
Bolts/Studs Shoulder engagement was too small Shear Shoe using two 3/8” dia bolts
5/16” Bolt thread necked down too far, shank not necked down, not compliant for thermal cycling
3/8” Studs necked down to root diameter, belleville washers
Torsion in long bolts during tightening, inaccurate tensioning
Studs with nuts used in place of long bolts, stud tensioner
Dual purpose bolts, combined tension and shear functions, tolerance issues, torque(tension) uncertainty
Loose fitting clearance holes for studs, separate shear shoes
Four 5/16" bolts @ 2500#, marginal friction to carry shear
Four 3/8" studs @ 5000#, doubling of preload
Thin washers under bolt heads at tee-ends 1/4" thick washer plate over Belleville washers at tee-ends
All defects contributing to original failure have been addressed
Feature Old Design New DesignInserts Keensert type, marginal thread engagement Taplok type, thread engagement > 0.5"
Shimming Manually selected and inserted G10 shim stock Hysol/glass tape potting in boxes, mold released to permit thermal growth
Out-of-Plane Load Path
Wedged G10 blocks with pusher bolts Flags potted in boxes, boxes bolted to hub disks
Torque Collar Two piece collar bolted directly to hub. Wet lay-up 0.25” thick Hysol RE2039 & HD3561. Holes in collar for epoxy outflow to enhance adhesion.
Three piece collar with sliding contact with hub for torsion-only connection. Wet lay-up 0.180” thick Hysol E-120HP (improved adhesive strength). Serrations in collar to enhance adhesion.
Joint Resistance Measurement
10A Biddle measurement via connection to two half flags on disassembled joint, 1µΩ resolution
200A precision measurement using voltage taps in situ, ≈ 20x enhanced resolution
CONCLUSIONS•New Design Corrects All Defects Associated with Original Design
•New Design Has Sufficient Margins at 6kG
• Follow-on Activities Will Increase Confidence
Mechanical Prototype Testing
Electrical Prototype Testing
Instrumentation During Commissioning and Operations
- voltage drop (200A maintenance and real-time) system- other (temperature, strain, displacement)